Skip to content
OpenSmartRoute

LLMs1 min read

SurveyAgent-HKA: Multi-Agent Framework for Scientific Survey Generation

SurveyAgent-HKA is a new multi-agent framework that generates scientific surveys by combining LLMs with human knowledge augmentation. Experiments demonstrate improved citation quality, structural consistency, and content quality compared to baseline approaches.

By OpenSmartRoute editorial · written through the router by writer-small

From arXiv cs.CL - “SurveyAgent-HKA: A multi-agent framework for scientific survey generation with LLMs and human knowledge augmentation”

The paper introduces SurveyAgent-HKA, a multi-agent framework designed to improve scientific survey generation. The framework addresses limitations of single-source LLM approaches by incorporating knowledge from published surveys and peer-review comments. It decomposes survey generation into distinct agent-managed subtasks. The initial outline is constructed by retrieving relevant papers from multiple sources and identifying key topics through clustering. This outline is then refined using outlines from related human-written surveys. Subsequently, topic-focused papers are retrieved and re-ranked to support drafting a well-grounded survey. Finally, the framework incorporates common issues identified in peer-review comments to guide revisions and finalize the survey. Experiments on two domains show that the approach outperforms mainstream baselines in citation quality, structural consistency, and content quality. The framework is also efficient in both time and cost.

Key details include the use of LLM-powered agents for each sub-task and the incorporation of data from published surveys and peer-review comments. The system retrieves relevant papers from multiple sources and identifies key topics through clustering to construct an initial outline. The refined outline is then used to select and draft the survey content. The framework leverages expert feedback from peer-review comments to guide revisions and ensure high-quality output.

This approach offers a practical solution for AI-assisted scientific writing applications. The efficiency of the framework, both in terms of time and cost, makes it suitable for broader applications. The framework’s architecture allows for modularity and potential integration with existing research workflows.

Source: https://arxiv.org/abs/2609.05938

Published Sep 9, 2026 · updated Sep 9, 2026 · 239 words

Keep reading

Related posts

More in LLMs

LLMs1 min read

CONDUIT: KV Cache Reuse Policy for VLMs

CONDUIT is a training-free refresh policy for KV caches in vision-language models, unifying single and multi-image reuse. It achieves high accuracy and speed improvements with a small refresh budget, outperforming budgeted methods on latency.

LLMs1 min read

Document-Level MT Evaluation Shows Statistical Equivalence

Research found that document-level machine translation evaluation, presenting full documents to annotators, yields statistically equivalent scores and rankings compared to segment-level evaluations. This suggests current document-level systems and associated metrics may not be accurately measuring intended aspects of translation quality.

LLMs1 min read

Dynamic Lagging Improves Simultaneous Translation with LLMs

Researchers introduced a sentence-trained LLM approach for cascaded simultaneous translation, utilizing prefix-aware fine-tuning to achieve flicker-free performance. This method, based on Qwen3-8B, demonstrates superior quality-latency tradeoffs compared to traditional wait-$k$ and target-suffix-deletion mechanisms.